Smart Grid Seminar

In the first part of this seminar, we describe how an aggregator coordinates a fleet of electric vehicles (EVs) to participate in the electricity markets and provide frequency regulation service to an independent system operator. Since the aggregate capacity comes from many EVs instead of a single source, the challenge is how to efficiently aggregate the small and uncertain capacity from the EVs and determine the bid of the aggregator. We formulate the problem as a stochastic program. Our formulation incorporates risk management using the conditional value at risk.  Efficient algorithms are proposed to tackle the formulated problem. In the second part of this seminar, we present a problem of load scheduling and energy trading in systems with high penetration of renewable energy resources. We adopt approximate dynamic programming to schedule the operation of different types of appliances including must-run and controllable appliances. We assume that users can sell their excess power generation to other users or to the utility company. Since it is more profitable for users to trade energy with other users locally, users with excess generation compete with each other to sell their respective extra power to their neighbours. A game theoretic approach is adopted to model the interaction between users with excess generation.

Joint work with Enxin Yao, Pedram Samadi, and Robert Schober.